| import os |
| import sys |
| import random |
| import torch |
| import numpy as np |
| import torch.nn.functional as F |
| from argparse import ArgumentParser |
| from core.registry import register_method |
| from core.base_method import BaseMethod |
|
|
| sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '../../Compact3DGS'))) |
| from utils.loss_utils import l1_loss, ssim |
| from gaussian_renderer import render as native_render |
| from scene import Scene, GaussianModel |
| from arguments import ModelParams, PipelineParams, OptimizationParams |
|
|
| @register_method("compact3dgs") |
| class Compact3DGSWrapper(BaseMethod): |
| def __init__(self, dataset_config, hyperparams): |
| self.parser = ArgumentParser() |
| self.lp = ModelParams(self.parser) |
| self.op = OptimizationParams(self.parser) |
| self.pp = PipelineParams(self.parser) |
| self.args = self.parser.parse_args([]) |
| self.args.source_path = dataset_config["source_path"] |
| self.args.model_path = dataset_config["model_path"] |
| self.args.eval = True |
| self.args.resolution = dataset_config.get("resolution", 1) |
| self.track_decoupling = hyperparams.get("track_decoupling", False) |
| self.dataset = self.lp.extract(self.args) |
| self.opt = self.op.extract(self.args) |
| self.pipe = self.pp.extract(self.args) |
| self.gaussians = GaussianModel(self.dataset) |
| self.scene = Scene(self.dataset, self.gaussians) |
| self.gaussians.training_setup(self.opt) |
| bg_color = [1, 1, 1] if self.dataset.white_background else [0, 0, 0] |
| self.background = torch.tensor(bg_color, dtype=torch.float32, device="cuda") |
| self.viewpoint_stack = self.scene.getTrainCameras().copy() |
| self.last_n_gaussians = len(self.gaussians.get_xyz) |
|
|
| def train_iteration(self, step): |
| self.gaussians.update_learning_rate(step) |
| if step % 1000 == 0: |
| self.gaussians.oneupSHdegree() |
| if not self.viewpoint_stack: |
| self.viewpoint_stack = self.scene.getTrainCameras().copy() |
| viewpoint_cam = self.viewpoint_stack.pop(random.randint(0, len(self.viewpoint_stack) - 1)) |
| |
| rvq_iter_flag = step > self.opt.rvq_iter |
| render_pkg = native_render(viewpoint_cam, self.gaussians, self.pipe, self.background, itr=step, rvq_iter=rvq_iter_flag) |
| image = render_pkg["render"] |
| viewspace_point_tensor = render_pkg["viewspace_points"] |
| visibility_filter = render_pkg["visibility_filter"] |
| radii = render_pkg["radii"] |
| gt_image = viewpoint_cam.original_image.cuda() |
| |
| Ll1 = l1_loss(image, gt_image) |
| ssim_value = ssim(image, gt_image) |
| loss_target = (1.0 - self.opt.lambda_dssim) * Ll1 + self.opt.lambda_dssim * (1.0 - ssim_value) |
| loss_parasitic = self.opt.lambda_mask * torch.mean(torch.sigmoid(self.gaussians._mask)) |
| loss = loss_target + loss_parasitic |
|
|
| stats = {} |
| grad_cos_sim = 0.0 |
| parasitic_ratio = 0.0 |
|
|
| if self.track_decoupling and step % 100 == 0: |
| param_groups_map = { |
| "spatial": [self.gaussians._xyz], |
| "geometry": [self.gaussians._scaling, self.gaussians._rotation], |
| "opacity": [self.gaussians._opacity, self.gaussians._mask], |
| "appearance": [self.gaussians.recolor.params, self.gaussians.mlp_head.params] |
| } |
| |
| self.gaussians.optimizer.zero_grad(set_to_none=True) |
| self.gaussians.optimizer_net.zero_grad(set_to_none=True) |
| loss_target.backward(retain_graph=True) |
| grads_target = {} |
| for g_name, params in param_groups_map.items(): |
| opt = self.gaussians.optimizer_net if g_name == "appearance" else self.gaussians.optimizer |
| valid_grads = [] |
| for p in params: |
| if p.grad is not None: |
| state = opt.state.get(p, {}) |
| v_t = state.get("exp_avg_sq", torch.ones_like(p)) |
| lr = 0.001 |
| for pg in opt.param_groups: |
| if id(p) in [id(param) for param in pg["params"]]: |
| lr = pg["lr"] |
| break |
| u = (lr / (torch.sqrt(v_t) + 1e-8)) * p.grad.clone() |
| valid_grads.append(u.reshape(-1)) |
| grads_target[g_name] = torch.cat(valid_grads) if valid_grads else torch.empty(0, device="cuda") |
|
|
| self.gaussians.optimizer.zero_grad(set_to_none=True) |
| self.gaussians.optimizer_net.zero_grad(set_to_none=True) |
| loss_parasitic.backward(retain_graph=True) |
| grads_parasitic = {} |
| for g_name, params in param_groups_map.items(): |
| opt = self.gaussians.optimizer_net if g_name == "appearance" else self.gaussians.optimizer |
| valid_grads = [] |
| for p in params: |
| if p.grad is not None: |
| state = opt.state.get(p, {}) |
| v_t = state.get("exp_avg_sq", torch.ones_like(p)) |
| lr = 0.001 |
| for pg in opt.param_groups: |
| if id(p) in [id(param) for param in pg["params"]]: |
| lr = pg["lr"] |
| break |
| u = (lr / (torch.sqrt(v_t) + 1e-8)) * p.grad.clone() |
| valid_grads.append(u.reshape(-1)) |
| grads_parasitic[g_name] = torch.cat(valid_grads) if valid_grads else torch.empty(0, device="cuda") |
|
|
| for g_name in param_groups_map: |
| gt = grads_target.get(g_name) |
| gp = grads_parasitic.get(g_name) |
| if gt is not None and gp is not None and gt.numel() > 0 and gp.numel() > 0 and gt.norm() > 0 and gp.norm() > 0: |
| cos = float(F.cosine_similarity(gt.unsqueeze(0), gp.unsqueeze(0))) |
| r = float(gp.norm() / (gt.norm() + gp.norm() + 1e-7)) |
| ti = r * max(0.0, -cos) |
| else: |
| ti = 0.0 |
| stats[f"sti_{g_name}"] = ti |
| |
| flat_gt = torch.cat([g for g in grads_target.values() if g.numel() > 0]) |
| flat_gp = torch.cat([g for g in grads_parasitic.values() if g.numel() > 0]) |
| if flat_gt.numel() > 0 and flat_gp.numel() > 0 and flat_gt.norm() > 0 and flat_gp.norm() > 0: |
| grad_cos_sim = float(F.cosine_similarity(flat_gt.unsqueeze(0), flat_gp.unsqueeze(0))) |
| parasitic_ratio = float(flat_gp.norm() / (flat_gt.norm() + 1e-7)) |
|
|
| self.gaussians.optimizer.zero_grad(set_to_none=True) |
| self.gaussians.optimizer_net.zero_grad(set_to_none=True) |
| loss.backward() |
| else: |
| loss.backward() |
|
|
| with torch.no_grad(): |
| if step < self.opt.densify_until_iter: |
| self.gaussians.max_radii2D[visibility_filter] = torch.max(self.gaussians.max_radii2D[visibility_filter], radii[visibility_filter]) |
| self.gaussians.add_densification_stats(viewspace_point_tensor, visibility_filter) |
| if step > self.opt.densify_from_iter and step % self.opt.densification_interval == 0: |
| size_threshold = 20 if step > self.opt.opacity_reset_interval else None |
| self.gaussians.densify_and_prune(self.opt.densify_grad_threshold, 0.005, self.scene.cameras_extent, size_threshold) |
| if step % self.opt.opacity_reset_interval == 0 or (self.dataset.white_background and step == self.opt.densify_from_iter): |
| self.gaussians.reset_opacity() |
| else: |
| if step % self.opt.mask_prune_iter == 0: |
| self.gaussians.mask_prune() |
|
|
| if step < self.opt.iterations: |
| self.gaussians.optimizer.step() |
| self.gaussians.optimizer.zero_grad(set_to_none=True) |
| self.gaussians.optimizer_net.step() |
| self.gaussians.optimizer_net.zero_grad(set_to_none=True) |
| self.gaussians.scheduler_net.step() |
|
|
| num_gaussians = self.gaussians.get_xyz.shape[0] |
| metrics = { |
| "loss": float(loss), |
| "loss_l1": float(loss_target), |
| "loss_ssim": float(loss_parasitic), |
| "num_gaussians": int(num_gaussians), |
| "delta_N": int(num_gaussians - self.last_n_gaussians), |
| "peak_vram_GB": float(torch.cuda.max_memory_allocated() / (1024 ** 3)), |
| "grad_cos_sim": float(grad_cos_sim), |
| "parasitic_ratio": float(parasitic_ratio) |
| } |
| metrics.update(stats) |
| self.last_n_gaussians = num_gaussians |
| |
| histograms = {} |
| if step % 1000 == 0: |
| histograms["opacity"] = torch.sigmoid(self.gaussians._opacity).clone().detach() |
| scales = self.gaussians.get_scaling.clone().detach() |
| histograms["scaling"] = scales |
| scales_2d = scales[:, :2] if scales.shape[1] >= 2 else scales |
| gamma = scales_2d.max(dim=-1)[0] / (scales_2d.min(dim=-1)[0] + 1e-7) |
| histograms["anisotropy"] = gamma |
| histograms["mask_activation"] = torch.sigmoid(self.gaussians._mask).clone().detach() |
|
|
| return metrics, histograms |
| |
| def render(self, camera): |
| with torch.no_grad(): |
| render_pkg = native_render(camera, self.gaussians, self.pipe, self.background, itr=-1, rvq_iter=True) |
| return {"image": render_pkg["render"], "depth": render_pkg.get("depth", None)} |
|
|
| def save(self, save_dir, step): |
| comp = True |
| store_npz = True |
| if step == self.opt.iterations: |
| self.gaussians.final_prune(compress=comp) |
| self.gaussians.precompute() |
| self.scene.save(step, compress=comp, store=store_npz) |
|
|
| def load(self, model_path, iteration): |
| load_path = os.path.join(model_path, 'point_cloud', f'iteration_{iteration}', 'point_cloud') |
| self.gaussians.load_model(load_path) |
| self.gaussians.precompute() |
|
|
| def get_spatial_centers(self): |
| return self.gaussians._xyz |
|
|
| def compute_physical_metrics(self, cameras=None): |
| metrics = {} |
| with torch.no_grad(): |
| scales = self.gaussians.get_scaling |
| scales_2d = scales[:, :2] if scales.dim() > 1 and scales.shape[1] >= 2 else scales.unsqueeze(-1).expand(-1, 2) |
| max_S, _ = torch.max(scales_2d, dim=1) |
| min_S, _ = torch.min(scales_2d, dim=1) |
| gamma = max_S / (min_S + 1e-7) |
| |
| metrics["gamma_median"] = float(torch.median(gamma)) |
| metrics["gamma_90th_percentile"] = float(torch.quantile(gamma, 0.90)) |
| metrics["scale_mean"] = float(torch.mean(scales_2d)) |
| metrics["alpha_mean"] = float(torch.mean(torch.sigmoid(self.gaussians._opacity))) |
| metrics["mask_mean"] = float(torch.mean(torch.sigmoid(self.gaussians._mask))) |
| |
| if cameras is not None and len(cameras) > 0: |
| view_dirs = [] |
| for c in cameras: |
| view_dirs.append(c.world_view_transform[:3, 2].tolist()) |
| view_dirs = F.normalize(torch.tensor(view_dirs, dtype=torch.float32, device="cuda"), dim=1) |
| rots = F.normalize(self.gaussians.get_rotation.clone(), dim=1) |
| w, x, y, z = rots.unbind(dim=-1) |
| normals = F.normalize(torch.stack([2*(x*z + w*y), 2*(y*z - w*x), 1-2*(x*x + y*y)], dim=-1), dim=1) |
| max_cos, _ = torch.max(torch.abs(torch.matmul(normals, view_dirs.T)), dim=1) |
| metrics["billboard_bias_ratio"] = float((max_cos > 0.90).float().mean()) |
| return metrics |
|
|
| def evaluate_spatial_field(self, query_points: torch.Tensor, cameras=None) -> torch.Tensor: |
| with torch.no_grad(): |
| V = query_points.shape[0] |
| densities = torch.zeros(V, device="cuda") |
| xyz = self.gaussians._xyz |
| opacities = torch.sigmoid(self.gaussians._opacity).squeeze() * torch.sigmoid(self.gaussians._mask).squeeze() |
| scales = self.gaussians.get_scaling |
| sigma_sq = (scales[:, :2].max(dim=1)[0].pow(2)) if scales.shape[1] >= 2 else scales.squeeze().pow(2) |
| |
| N_gaussians = xyz.shape[0] |
| chunk_size = max(1, 30_000_000 // (N_gaussians + 1)) |
| for i in range(0, V, chunk_size): |
| end = min(i + chunk_size, V) |
| dist_sq = torch.cdist(query_points[i:end], xyz, p=2).pow(2) |
| weights = torch.exp(-0.5 * dist_sq / (sigma_sq.unsqueeze(0) + 1e-7)) |
| densities[i:end] = torch.sum(weights * opacities.unsqueeze(0), dim=1) |
| return densities |
|
|
